The right hemisphere (RH) processes language in a qualitatively different manner than the left hemisphere (LH). Most evidence comes from the subtle deficits in RH-damaged patients, and from visual hemifield methods that show distinct patterns of processing for words directed to the right visual field-LH (rvf-LH) or to the left visual field-RH (lvf-RH). Few studies have explicitly examined RH language processing with neuroimaging. We propose a 3-year program of 19 experiments to investigate semantic processing in healthy subjects using whole-brain functional Magnetic Resonance Imaging (fMRI) and visual hemifield methods. To date fMRI has rarely been used to study hemispheric differences in language processing, and compared to behavioral methods will provide more direct and anatomically precise information. The project extends preliminary experiments demonstrating RH-specific fMRI signal in semantic tasks. Comparison of fMRI and visual hemifield methods will help pinpoint, functionally and anatomically, the loci of effects observed in hemifield experiments. Both methods will be used to test the following predictions: 1) When subjects respond to lvf-RH words, greater fMRI signal will be observed in RH homologues to various language-associated areas than when subjects respond to rvf-LH words. 2) Hemispheric differences are equivalent across sex, but not across handedness. 3) fMRI signal will increase in the RH for tasks that yield a lvf-RH advantage, and increase in the LH for tasks that yield a rvf-LH advantage providing converging evidence - or not. Specifically: 4) Recognize that target words are related to ambiguous words imbedded in sentences biased against the target meaning. 5) Generate "unusual" actions distantly associated with nouns. 6) Solve or make solution decisions for Compound Remote Associate Problems. 7) Judge relatedness of multiple distantly related words. 8) Judge the relative typicality rather than category membership of exemplar words. 9) Answer questions about predictive inferences, but not coherence inferences. Predictions are derived from the Bilateral Activation, Integration, and Selection framework of semantic processing, which contends that the RH engages in coarser coding than the LH for each of these component processes, which are related to specific cortical areas (posterior temporal, anterior temporal, and inferior frontal, for semantic activation, integration, and selection, respectively).